Scary Statistics Made Simple with Thomas DeMarco

Dan Swartwout: We are lucky enough to be joined by Mor Steam's principal statistician, Thomas DeMarco. Thomas, welcome to the show.

Thomas DeMarco: Thank you, Dan. Great to be here.

Dan Swartwout: We are so excited to have you here. And we have done a lot of shows recently on the people side of Lean Leadership, and it hasn't been as heavy in the statistics as some of our earlier episodes. So I wanted to do kind of a focus shift and bring on you, Thomas, our principal statistician, to learn a little more about you, what you do at more steam, and why statistics are so important in continuous improvement and operational excellence. So we're excited to have you here today, Thomas.

Thomas DeMarco: Thank you, Dan. It's great to be here. Yeah. Um, and happy to talk about those things.

Dan Swartwout: Great. Now you got both your BS and, and Ms. In mathematical sciences at Clemson with a concentration in applied statistics. What drew you toward the applied side specifically rather than pure math?

Thomas DeMarco: Um, I've always had a pull towards kind of building things, um, engineering. Um, most of my family are either engineers or dentists. A little interesting mix, but they're, they're doers who, who create things. And, um, But I always had kind of what I was told, and I was pretty good at math, so I was wondering kind of what could bridge that field. I, I started, you know, my college career and, you know, looking at jobs in the actuarial field more that pure math kind of side. Um, got a taste for that and then also got a taste for kind of a applied statistics role afterwards. And, you know, I didn't know at the time when I was, you know, uh, when I was starting, but I learned that this is where you can bridge that gap of knowing all the pure math, doing, having the pure math and working, um, with people who are doing the building and the creating. And so I really love working in, you know, uh, engineers, chemists, operators, backyards. You know, they're out there making the world a better place, and I'm just helping them do it. So I, you know, that's, that's what drew me in for sure. To the applied side.

Dan Swartwout: As an aside, one of the things I try to work into the show as much as I can is college football. Were you at Clemson during one of the national championship runs?

Thomas DeMarco: Uh, only both of them.

Dan Swartwout: Oh, whoa, whoa, whoa, whoa, whoa. You weren't with the one in 1980

Thomas DeMarco: with Danny Ford, were you? I, you know, I was not. I was not in the 80s. No. That was our first run, so. Very good football knowledge. Dan knows ball, that's for sure. I, I was Clemson's last, uh, graduating class or last freshman class, however you look at it, that joined Clemson, not because it was a football school. When I applied to Clemson, we were not good at football. And then my freshman year we randomly go to a national championship and lose later to win two more times and lose another time. All right, So I, I was spoiled, but it was not why I went to Clemson. But I definitely fell in love with it.

Dan Swartwout: All right.

Thomas DeMarco: All right.

Dan Swartwout: Now, before more steam, you spent four plus years as an industrial statistician at Eastman Chemical. What does a statistician actually do inside a big chemical company on a day to day basis?

Thomas DeMarco: I always like to say I was, you know, Eastman was a really big company manufacturing all kinds of things. Things from chemicals to paints to, um, plastics to, you know, a lot of the raw materials you use in everything day to day. The screens you and I are looking at have raw materials that Eastman probably made. And with that large company, I was kind of just a stats consultant to the company. So my day to day depended on what they were doing. Right. You know, that's the applied side that I love to get into is whatever the engineers are doing. I'm helping them, um, understand the story of their process that their data is trying to tell them. So whether that's teaching them statistics or just doing some of the more advanced statistics for them and getting them understanding. Um, because you know, if I just understand it, then it's useless because I'm not the doer, I'm not the one turning the knobs. Um, and so it's a balance of definitely teaching and then doing the more advanced stats that um, you know, whatever's cutting edge at the time.

Dan Swartwout: Now the title principal statistician carries a lot of weight. Thomas, what does it actually mean in practice? What are you responsible for here at Morris Team that a typical statistician role somewhere else might not cover?

Thomas DeMarco: As a statistician at Morris Team, I'm, um, definitely responsible for, you know, or I'm really the subject matter expert in all things statistics. So, um, the things I touch, whether it be marketing or elearning or building some of our great software such as engine room or traction, you know, the statistics that goes on in there is really kind of backed by me. And how I get, you know, make sure I'm in the right place is I work a lot with our great, uh, Morristein Client Services, a phenomenal group of statisticians and practitioners who have many more years of experience than I do. Um, and working with them sometimes to make sure I've got my head in the right place. And then, you know, actually working with the different, uh, you know, areas here at Moresteam to make sure that our stats are sound and that we're going in the right direction to help our customers.

Dan Swartwout: One of the things I heard, Thomas, is that before you joined More Steam or looked to work at More Steam, you really liked engine room software and that was something you wanted to work with. Could you tell us a little bit about that, about your initial exposure to engine room, what you like about it, why you wanted to work with, uh, it, and what you continue to do with engine room?

Thomas DeMarco: Stats tools in general have a common relationship between ease of use and statistical power. Um, it's not a direct one to one shift where you increase statistical power and you lose ease of use, but there is a strong correlation. And wherever you find your statistical power at in the range of really basic to extreme advanced stats, there are things that you can do when building a software and designing a software to try to push that needle of ease of use as far as it can go. And engine room does a phenomenal job at doing that, at uh, making sure the user understands what they're doing, gets the output they need and gets, you know, the layman's terms of the statistics that they're working on. And you know, if you need the assist of what is this p value mean, we give you that assist. Or if you are a more advanced statistical user, you don't need that. So we're not going to throw it in your face, but we make sure we have the ease of use for, you know, our yellow belts who are kind of learning the tools from the first time, seeing all these random Greek letters and dead statisticians, names for tools that, you know, they don't know what a moods M median test is. Right. So it's confusing. Um, you know, and there are a lot of big words in statistics. Um, you know, I'm a math major or I was a math major, so, you know, I definitely am scared of big words myself. That's not my expertise. And we're trying to improve that ease of use as best we can and reach the statistical power of our customers and where they need it because the fundamentals are really what they need, um, for a lot of these, um, lean six sigma black belts and um, the, you know, quality improvement engineers we work for.

Dan Swartwout: Sometimes people tend to get stuck when it comes to the more advanced statistics. What's the cost of that? When you get stuck on the more advanced statistics, or maybe you're More scared of the advanced statistics because of that ease of use as you're going through your process improvement, continuous improvement projects.

Thomas DeMarco: Yeah, no, that's a fantastic question. And I think, you know, with the exponential growth of technology as it goes, um, this is a common problem that's not going to go away. Um, where and the way I think that we're positioning ourselves to combat that is by really teaching the core fundamentals better for these practitioners. You know, you can jump to the m most advanced statistics at the click of a button. Nowadays you can click a button and your data set that you downloaded with two clicks of a button, uh, is now being run in a neural network. And that black box gave you an answer to a question that you don't even know what it is. Um, and so with that, um, at our fingertips, sometimes we can jump to the shiny new statistics far too quickly without kind of getting the fundamentals. While boring, the m fundamentals are what can get us to those advanced stats tools and use them effectively, be able to defend them and apply them appropriately. Um, but I truly believe that the fundamental stats, especially the core, um, aspects of engine room and our elearning and our lean six Sigma, um, training are what can get you, if you have the full understanding, the full ability to defend it, the full ability to apply it gets you to improve your process to 95% of what you're imagining. And from there, if you have those fundamentals and the time to then learn the more advanced stats, um, really branching out to the new and powerful things that are phenomenal achievements in statistics nowadays. You, um, can then bump that up to 96, 97, 98%. But if you skip ahead then you're just left with confusion and um, again, not even knowing the question that we were trying to answer with these advanced models that you can run at the click of a button. So our my answer is sort of I kind of combat that desire and reach for more advanced stat tools with. Let's make sure you have the fundamentals first. Um, because I think in society that's where we're drifting away from. Um, and you know, once you have that, you learn. Wow. I can, I can kind of manage my process just from here, just with, with these basic tools. So I sound like when I'm answering that, I sound like, I always think about the Buffalo Wild Wings commercial with like the Buffalo and he's like the kids nowadays in March Madness don't know the fundamentals. And then he like boxes out people through a window, things like that, you know, and that's what. And that's truly how I'm thinking of it here. If you don't. And I've definitely watched a few March Madness games and yelled at them to box out because, you know, that's not cool anymore. Right? That's what wins games. Right.

Dan Swartwout: I think about with baseball. It's baseball season now, too. I think about the way I was taught to field a ground ball by, you know, getting down on one knee, raising the other leg up so that my whole body is in front of the ground ball, that if I don't get it with my glove, some part of my body blocks the ground ball. So I watch baseball now, and I'm like, that's a little different than what I was taught.

Thomas DeMarco: Yeah, no. Uh, I played baseball for a few years, and I definitely was getting yelled at for just trying to do glove only and not getting my body behind it. And that's why I often was running after balls that went over my head.

Dan Swartwout: So we're talking about some of these advanced statistical concepts. I have seen you, Thomas, teach. I've seen you do webinars, and you do a really good job of making those advanced statistical concepts more approachable, more down to earth, more actionable. What does that look like in practice? And how do you do that without dumbing it down, so to speak?

Thomas DeMarco: Statistics really all comes from reason. Um, and as a teacher or, um, as anyone teaching statistics, I think their main job is just to bridge the gap between the statistical concept and the reason that everybody understands. Um, you know, because again, we have Greek letters, things named after statisticians, um, that are confusing and can kind of mask and make it seem like, oh, these are really complicated things that you can only understand if you, you know, really grind through the mathematics instead of just bridging it with the reason M. And so, you know, recently I taught a DOE fundamentals course, and the best way I know how to teach orthogonality, which is a really big word. Um, that is a big word.

Dan Swartwout: I don't. I have no idea what that word means. Just. So go on.

Thomas DeMarco: Exactly. So it's a fundamental, uh, principle of designed experiments. When you're running an experiment, change turning knobs to see how it impacts your response, your output. You know, what happens if I turn up the temperature in this oven? What does that do to my cookies that I'm baking? That's, you know, our cookies.

Dan Swartwout: I know that word, Thomas. Cookies.

Thomas DeMarco: I know that word.

Dan Swartwout: So now I'm dialed in. I'm totally dialed in.

Thomas DeMarco: We're dialed in. We're dialed In. Well, now imagine that you're actually just in a room, maybe your room, that the office that you're sitting in has two light switches. Mine has two light switches here. Um, and you go in and both the light switches are down and the lights are off. And you just want to turn the lights on. Right. So often what do people do? You, you flip both switches on, right?

Dan Swartwout: Mhm.

Thomas DeMarco: And when you flip both switches on, then usually the lights will come on. And if you're in a big meeting room, you might have a lot of lights turn on. And so, you know, you got to what you were after, you wanted the lights on, but if someone were to ask you, hey, can you just turn on the front lights or just turn on the back lights, we're going to do a presentation at the front. When you went from both of those light switches off to both of those light switches on, you don't know exactly what light switch does what. You know how to get from all off to all on, but you don't know independently what each light switch does. And that can happen a lot in experimentation where we change too many things at once and maybe we get to what we're after. And that's great. We, we can sometimes, um, take some risks and take some shortcuts to get to what we want. But we do lack that understanding of what each light switch does. There's a chance that one light switch does all the lights and the other light switch is a blank. Right? That's how mine works in my office. Because there could be a fan in here one day, but it's just an empty blank, uh, light switch. Or it could be that one light switch turns on the front half, one light switch turns on the back half. And the only way you're going to know that is if you independently switch those light switches on and off. And that once you do that, once you have both off, both on, one off, one on, um, you'll gain what's called orthogonality, experimentation, which is just balance. You independently change those two light switches so you know what each of them do. Right? But if you flip both of them on and both of them off, you'll never know what they do independently and you won't have orthogonality. So that's what that big word means, is just figure out what each switch does.

Dan Swartwout: Thomas, you answered my question in a way I wasn't expecting. You answered my question about bringing these concepts down to Earth and making them actionable by explaining the concept to me. And that was even better. That was even better than Just explaining what you did by showing me how you did it. So now I understand. When you're looking at these, uh, our products, when you're looking at more Steam's products and you're thinking about the end user, because every time we, we want to make sure everything is 100% scientifically sound, statistically sound, and we come to you to verify that if need be. What are you looking at as it relates to the end user? Whether it's yellow belt training, master black belt training, engine room. What are you thinking about for the end user when you're looking at statistics here at Moresteam.

Thomas DeMarco: Gotcha. Yeah. So trying to put uh, myself in the doer's shoes that were really um, making these products for, you know, um, I think statisticians, uh, without, you know, the boots on the ground, it's uh, we're useless. Right. Um, my job can never be completely independent. Right. I need someone out there doing the work. Right. And um, so who we definitely make our products for are those doers. You know, there are some super advanced statistics out there done by super advanced statisticians. Um, but our products are really for the end user. Like you said, the yellow belts, black belts, master black belts, who are, they aren't statisticians, they're um, they're doers, they're managers, they're operators, they're scientists, chemists, um, and they're out there making these changes and making these decisions and we're arming them with the capability to understand their process better by giving them the tools. They have data and the data is trying to tell them the story of their process and we are giving them the tools to understand that story. That's who I'm thinking about. I'm m thinking about those um, belts and those quality engineers, those scientists, operators, managers who didn't go and get statistics degrees. So they need that level of understanding. How do we get them that understanding? We have the elearning, we have the training, we have the blended programs that can get them there and then we also have very high ease of use software that has the appropriate statistical power for that.

Dan Swartwout: Thomas, this has been a fantastic interview. Thank you so much. Uh, I want to close with this because I know I'm going to have you on the show again for some more in depth discussions about some really interesting statistical concepts. But, but right now I just want to know what are you most excited to dig into here at More Steam and, and what is on your radar that you think our listeners should be paying attention to?

Thomas DeMarco: You know, that's a great question. Um, you know, getting to look at what we're doing here at Morristein for all those practitioners we're talking about, um, and the biggest thing I talked about before, the, the ease of use versus statistical power. And I think we're still really, um, you know, we're building statistical power when needed. If we find a gap that customer needs, um, I challenge people listening to this to bring that up to us and see how fast we implement it.

Dan Swartwout: Right.

Thomas DeMarco: Um, so we're still working on those things as they come up, um, making sure to keep an eye on what our customers need. But then for ease of use, that's not something really a customer ever asked for. That's something we have to create. And um, in Engine Room 5.0, this is kind of something we've been talking about that's coming up. And in that we have something that I think, uh, really I've used a lot of statistical tools in my time in school and then, you know, at Eastman, um, as a stats consultant. And one thing that you know, we're trying to add is something that's never really been a part of the more, uh, powerful stats tools, which is collaboration, um, and collaboration is going to allow us, or allow the users to get that higher ease of use by being able to work alongside their mentors, their subject matter experts, their team members, um, simultaneously in their same workspace that also has all that stats power and ease of use that we have in current engine room. And being able to collaborate there, you know, it not only speeds things up, but it improves understanding. Um, we're not sending files back and forth and losing, you know, what's going on, where, what's a test to what. We have it all in one kind of source of the truth. And I think that that collaboration is going to be huge, um, for the customers. Ease of use. So that's what I'm excited for. That's being worked on, um, and I'm getting the pleasure to test it out and guide its development.

Dan Swartwout: More Steam's enterprise process improvement platform includes the tools, training and software you need to transform your organization into a problem solving powerhouse. Join the ranks of Fortune 500 companies that trust Morristeam for their enterprise process improvement. Visit Moresteam.com to learn more. Also, I want to take just a moment to thank you for listening to Quality Time with More Steam. You can email the show@qualitytimeoresteam.com and if you're enjoying the show, please leave a rating or review on your favorite podcast service. Thomas, thank you. So much. I'm, um, looking forward to seeing you again in person at More Steam's Best Practices for Operational Excellence Conference in a couple months. This has been a tremendous conversation. We'll have you on again soon to really dig deep, but I really wanted to introduce you as our statistician to the listening audience to let them know just how ingrained statistics are here at Moore Steam, and of course, with Thomas DiMarco.

Thomas DeMarco: Well, thank you so much, Dan. It's been great talking with you. Always a pleasure.

Dan Swartwout: We'll be back in two weeks with another episode of Quality Time with More Steam. Thanks again.